GEORGE ELBERT KIMBALL: OPERATIONS RESEARCH INNOVATOR 1906-1967
by
Philip M. Morse
George
Kimball was a generalist, capable of achieving outstanding
recognition in two fields of science and of leaving
his mark on other fields of human endeavor. Perhaps
his greatest contribution was the education and
inspiration he gave to many younger men, now working
in various fields of science and technology.
Kimball was born in Chicago in 1906 and grew up
in New Britain, Connecticut, displaying few remembered
signs of precocity and no marked preference for
science. He took all the Latin his local high school
could provide but his chemistry teacher was the
one who caught his interest. After a year at Exeter
Academy, he went on to Princeton, in 1924, urged
there by his father, who felt that there were too
many Yale graduates in Connecticut.
George was a fairly typical undergraduate at Princeton—he
was on the water-polo team—though not a typical
chemistry student of the time. He later claimed
he chose the chemistry program because it allowed
him to take as much physics and mathematics as chemistry,
and he wanted to learn all three subjects. By the
time Kimball received his bachelor's degree, in
1928, his interest had centered on quantum chemistry.
His abilities were such that the Chemistry Department
offered him one of its best graduate fellowships,
so he returned to Princeton, to work under Hugh
Taylor, to soak up more physics and mathematics
and, though yet a graduate student, to give a private
course in quantum mechanics to the faculty of the
Chemistry Department. E. Bright Wilson, who was
two years behind Kimball, remembers that "I
had an enormous respect for his knowledge and his
ability to explain things. He seemed to know everything,
and I think he really did. It was not at all that
he was boastful or a show-off—I used to seek
him out for enlightenment, and he always provided
it." Kimball's thesis for the doctoral degree,
granted in 1932, was on the quantum mechanics of
the recombination of hydrogen atoms.
He applied for a National Research Fellowship in
physics for 1932-1933 but missed out because he
was not well known to the physics fraternity. He
stayed on at Princeton as instructor and next year
applied for a National Research Fellowship in chemistry
and won, going to MIT for the two years 1933-1935.
Though he was officially assigned to the Chemistry
Department, he spent much of his time working with
Slater and others in the newly reconstituted Physics
Department at the Institute.
These were heady times. The new faculty was augmented
by a galaxy of postdoctoral fellows, and a new generation
of graduate students enlivened the scene. Among
this aerie of eaglets Kimball more than held his
own. In 1965 he wrote a reminiscence of those times
for the International Journal of Quantum Chemistry,
which says, in part:
The
group which inhabited the third floor of the Eastman
Laboratory sat at the feet of an academic trinity.
John Slater (then 33 years of age) was the Old Man,
with a long and illustrious career behind him. Philip
Morse was the junior member of the trinity (he and
I had been graduate students together at Princeton).
The third, and most spiritual member was Julius
Stratton, another old man (only a year younger than
Slater), who mystified everyone (except Bill Hanson)
by being more interested in Maxwell's equations
than in the Schroedinger equation. "The great
Depression was at its height (my first job after
I earned my Ph.D. paid the magnificent salary of
$900 a year). As a result the group of graduate
students and postdoctoral fellows with whom I worked
lived a sort of Vie de Boheme. The center of this
life was the third floor of the Eastman Laboratory,
where we shared office space. We spent our evenings
as well as our days there, but not always at our
work. There was a ping-pong table, and someone discovered
that the long, long corridors of M.I.T. made a wonderful
place to roller skate.
Every afternoon we had tea, served by Alice Hunter,
student in chemistry, who has since done me the
honor of becoming my wife. Those teas became a sort
of discussion group, led by Norbert Wiener, who
would argue violently on any subject, such as Chinese
grammar, or whether or not the number of palindromic
primes is infinite.
From time to time we would have a party. The most
famous of these was a theater party at which we
all had seats in the second balcony to see a D'Oyly-Carte
performance of The Gondoliers. Alice Hunter
brought her knitting, including a large ball of
bright orange yarn. During an intermission Ralph
Johnson, sitting beside her, picked up the yarn
and Satan (in the person of Bill Shockley) whispered
'Throw it.' Ralph did, all the way to the orchestra
pit. Someone tried to throw it back, but it only
got as far as the first balcony. From there it was
thrown back and forth until the whole theater was
festooned with orange yarn. Finally an ingenious
usher broke the yarn and carried the remains of
the ball up to the second balcony.
My office mate was George Shortley. He was putting
the finishing touches on Condon and Shortley's Theory
of Atomic Structure. It is interesting to note
that out of this rather small group, three of us,
Philip Morse, George Shortley and myself have since
served as presidents of the Operations Research
Society of America. During the summer of 1935 he
returned to Princeton, to work with Eyring. After
a year spent teaching physics at Hunter College,
he entered the Chemistry Department of Columbia
University as an assistant professor. With a few
intermissions he remained there until 1956, becoming
Professor of Chemistry in 1947.
The
five years 1936-1941 were productive ones. He published
nine papers on reaction rates and electrochemical
surface effects, he introduced and taught courses
in quantum chemistry, and he supervised graduate
student research.
Dr.
Isaac Asimov, a student of Kimball.
One
of his students, Dr. Isaac Asimov, writes: "I
had a lab course from Kimball in physical chemistry
and at one time was asked one question out of a number
of possible questions and drew a complete blank. I
got a zero. I came to him afterward and said that
the question I was asked was the only one of the alternatives
I couldn't answer perfectly and that a mark of zero
was not a true measure of the state of my knowledge.
He said, 'The time will come when you will be asked
a question, and it will be the only one of a number
of alternatives which you can answer perfectly. You
will then get a mark of one hundred and that will
not be a true measure of the state of your knowledge
either. But you will not complain then, will you?'
Very much against my will, I saw the justice of that
and subsided. I kept my zero—but I passed the
course."
In 1942, when I was asked by the Navy to organize
a group to analyze antisubmarine tactics, Kimball
was one of the first persons I recruited. Within the
year he became Deputy Director of the group, called
the Operations Research Group (ORG) during the war,
later called the Operations Evaluation Group, U.S.N.
It grew to number seventy-odd analysts by 1945. Kimball's
abilities were in daily use as an educator, as a universal
scientific encyclopedia, and as a deviser of simple
algorithms to solve tough problems quickly. His colleague
then and later, Arthur A. Brown, comments: "In
the ORG the initial work dealt with search and with
the optimum geometric patterns for the depth charge
bombing of German U-boats. In a very short space of
time the group was working on tactical patterns for
destroyer attacks, on the question of reliability
of aircraft sightings, and the related question of
whether or not to send out a destroyer force.
This brought us into the question of convoy protection
and convoy size, and into liaison with the Coastal
Command of the R.A.F. The work spread into the South
Atlantic and into the Pacific, in relation to our
own submarine offensive against the Japanese supply
lines; to our combat air patrols against attacks on
the Third and Fifth Fleet operations; and to defensive
tactics against Kamikaze attacks. By the end of the
war the group had a network of field operations and
a solid place in Washington's strategic councils.
George Kimball was in the midst of all this and he
contributed largely to it.
The day after Hiroshima, I, as Director of ORG, obtained
one of the first copies of the "Smythe Report"
to arrive in Washington. Within twenty-four hours
my Deputy, Kimball, and I briefed Admiral King and
his staff on the naval implications of the A-bomb;
a day later we briefed Secretary of the Navy Forrestal
and the joint Senate-House Naval Affairs Committee.
In the midst of all this, Kimball's work with Eyring
and John Walter, started ten years earlier, was completed
and the book Quantum Chemistry was published
in 1944. At the end of the war some of the Operations
Research Group decided to delay returning to their
peacetime positions long enough to record what had
been learned. Kimball and I wrote the volume Methods
of Operations Research; Bernard Koopman, who had
joined the group in 1943 wrote Search and Screening;
and Charles Sternhell and Alan Thorndike wrote a technical
history of Anti-Submarine Warfare in World War
II. All three volumes were initially classified.
It was not till 1951 that the first volume was declassified
and turned over to a commercial publisher; the other
two volumes were never declassified. The Methods
still is used as an introduction to the subject, and
is still referred to in the literature.
Kimball
was elected to the National Academy of Sciences
in 1954.
After
this, Kimball went to the Chemistry Department at
Columbia, to resume his research and teaching in
theoretical chemistry. That he was successful is
evidenced by the dozen papers he published on chemical
kinetics and on other subjects in chemical physics.
Honors began to come his way. He received the Presidential
Citation of Merit for his war work; he was elected
to the National Academy of Sciences in 1954. He
also retained his interest in the new field he had
helped pioneer during the war, operations research.
He continued his contacts with the Navy, acting
as consultant with the Operations Evaluation Group
and serving on various advisory panels on underwater
ordnance. When the Weapons Systems Evaluation Group
(WSEG) was formed in 1949, to carry out operations
analysis for the Joint Chiefs of Staff and the Secretary
of Defense, he contributed to its work, for a time
as consultant and then for a while as a full-time
member of the group. He also assisted in organizing
the NATO Advisory Panel on Operations Research.
Even during the war Kimball had become convinced
that operations research could be effectively applied
in industry and in the public sector. He was interested
in enlarging public awareness of its potentialities
and was active in organizing the Operations Research
Society of America, which was founded in 1952, with
Kimball as a member of the society's first council.
By 1964, when he was elected the society's president,
the society had about 5,000 members.
In the 1950s Kimball began to spend some time with
the operations research division of Arthur D. Little,
Inc., assisting in its consulting work for industry
and for the Navy. This work increasingly engrossed
his attention until, in 1956, he left Columbia and
came full time to A. D. Little, first as Science
Advisor and then, in 1961, as Vice President. When
asked, later, whether he missed teaching, he replied
that he was still teaching and that it was a greater
challenge to teach people who didn't want to learn
or didn't know they were learning.
Much of his work with A. D. Little dealt with applications
of theory to the specific problems of the client.
Most of this has of course not been published. A
partial list of his internal reports and notes indicates
that he initiated developments in dynamic programming,
decision theory, inventory, and reliability theory,
which others fed into the open literature later.
Kimball was never particularly interested in publication.
He would spend a great deal of time solving specific
problems of immediate importance, or in making clear
the underlying theory to clients or to classes,
but to establish priority by publication, with all
its drudgery of typescript, galley and page proof,
had less attraction for him than some new problem.
He always maintained that there was too much publication
anyway.
Arthur Brown reports Kimball's comment when Brown
said that someone should have pointed out that the
theory of search anticipated the basic principles
of information theory; that the probability of target
detection is just the entropy of the target distribution.
Kimball remarked that this was of course true, but
everyone had known about entropy for decades. In
his view all he had done was to apply known theory
in a context which needed theoretical clarification.
Kimball also did his part as citizen and parent.
John B. Lathrop, a neighbor and colleague at A.
D. Little, reports: "George had a strong sense
of responsibility to the community and gave it as
much time as he could. He spent many years as officer,
committeeman, or consultant for church, Boy Scouts,
and community. An example is his study of the growth
of the school population of his home town of Winchester,
Mass., done for the local School Committee. His
classification of the people of Winchester as 'old
families,' 'new families in old houses,' and 'new
families in new houses' and his tracing of the different
patterns of change and incidence of school-age children
in these groups was a model of useful statistical
analysis, forecasting, and clear understanding of
the phenomenon he studied. He concluded the report
with a basis for decision—the earliest, expected,
and latest dates when various school additions would
be overcrowded.
"George had strong interests—and really
was expert—in many fields; languages, naval
history, bridge, music, cooking. For years he made
almost all the family's gravy, and taught his children
how (he was convinced it took a chemist to make
a good gravy). There was a blackboard in the family
kitchen and frequently he would sit there over a
cocktail while dinner was under way, discussing
calculus or chemistry or physics with one or another
of the children. Three of the four children have
definite scientific leanings."
For his last several years, Kimball suffered from
serious cardiac illness. For the final year, at
least, he was in constant pain. Those who saw him
daily knew that this was so, but none were made
aware of it by his manner, his actions, or his words.
He continued to work actively on all projects and
in all the fields that interested him. His death,
in fact, came in the midst of his duties; when he
died, on December 6, 1967, he was in Pittsburgh
as a member of the Visiting Committee of the Carnegie-Mellon
University's Chemistry Department. At the time of
his death he was chairman of the Northeastern Section
of the American Chemical Society.
Kimball was a generalist—which doesn't seem
to rate the acclaim the specialist gets nowadays—but
his value to operations research, indeed to science,
lay in his universal interests.
Many of us would agree with Joseph Mayer's comment:
"George and Alice were delightful friends to
have and we enjoyed them immensely. I have always
thought that George was one of the most pleasant
companions of an evening, with whiskey and soda
after a good dinner. He was not particularly a person
who sparkled; he was just comfortably tolerant and
very intelligent and informative."
Kimball's style of work was rooted in his personality.
It was characterized by simplicity of thought and
method. Another characteristic was theoretical power
and depth. A third was a permanent adherence to
reality. He never liked the spinning of elaborate
webs of mathematics and he never liked to be too
far from actual data. He was sensitive to problems
of wording, emphasis, and timing in the presentation
of research results, but he was wholly uncompromising
in matters of principle. He set an example worth
following.
He could bring concepts from chemistry to bear on
inventory and marketing problems; he could devise
an abstruse mathematical algorithm to make a digital
computer produce random numbers as fast as was needed.
Everything he did had to be done well; if he couldn't
do it well he didn't do it. In fact, his uncompromising
standards kept him from publishing much good work,
because it wasn't in final, polished form. Many
of us wished that more of his lectures could have
reached a wider audience, but that was not his way.
He preferred to work directly with people, not via
the printed word. And this was in line with his
concentration on immediate problems, rather than
on abstract theory.
Kimball's style of work was rooted in his personality.
It was characterized by simplicity of thought and
method. Another characteristic was theoretical power
and depth. A third was a permanent adherence to
reality. He never liked the spinning of elaborate
webs of mathematics and he never liked to be too
far from actual data. He was sensitive to problems
of wording, emphasis, and timing in the presentation
of research results, but he was wholly uncompromising
in matters of principle. He set an example worth
following.
Abridged from Biographical Memoirs,
National Academy of Science,. Vol. 43 (New York:
Columbia University Press, 1973).
Unitarian
Note
Dr.
Kimball's activities included service as a trustee
and president of the Unitarian congregation in Hackensack,
New Jersey.
